Influence of atlas-based and patient dependent forward models in EEG source reconstruction

Abstract: Electroencephalography Source Imaging (ESI) techniques have become the most attractive alternative to support the estimation of neuronal activity through the mapping of electrical potentials measured over the scalp. It takes advantage of the low implementation cost, the high temporal resolution, and non-invasiveness in the patient. ESI techniques require a volumetric conductor model (commonly named Electroencephalography (EEG) Forward Model), including information about the physiological and geometrical properties of the head, and modeling the electromagnetic field propagation of the neuronal activity throughout the head tissues to reach the scalp. In this regard, the accuracy of ESI solutions depends partially on the capabilities of the forward model to correctly describe the structural information provided by a Magnetic Resonance Image (MRI). However, acquiring MRIs for generating personalized head models is expensive, slow, and in some cases unpractical. In this work, we investigate how the head model influences the source reconstruction based on EEG when progressively including different levels of prior structural information. Hence, we evaluate two approaches to enhance the model of brain structure in the EEG forward problem formulation. First, the incorporation of different brain tissue morphology, mainly, based on a Generic MRI, based on a target population Atlas, or based on a patient-specific MRI. Second, the variation of the tissue model complexity in the number of segmented brain layers. All the head models are build using the Finite Difference Reciprocity Method (FDRM). Model comparison is carried out under a Parametric Empirical Bayesian (PEB) framework using Event-Related Potentials (ERPs) taken from the studied population. Obtained results show that the more realistic and subject dependent model, the better performance of the ESI solution

• Autores: Céspedes Villar, Yohan Ricardo
• Fecha de publicación: 2017
• Institución: Universidad Nacional de Colombia
• Repositorio: Universidad Nacional de Colombia
• Idioma: spa
• OAI Identifier: oai:repositorio.unal.edu.co:unal/62118
• Acceso en línea: https://repositorio.unal.edu.co/handle/unal/62118; http://bdigital.unal.edu.co/61026/
• Palabra clave: 61 Ciencias médicas; Medicina / Medicine and health; 62 Ingeniería y operaciones afines / Engineering; EEG; Modelo directo; ESI; BMS; Electrónica médica; Electroencefalografía; EEG; Forward model; ESI; BMS; Electroencephalography; Medical electronics
• Rights: openAccess
• License: Atribución-NoComercial 4.0 Internacional